import numpy as np
import matplotlib.pyplot as ppl
import matplotlib.cm as cm
# Custom modules
import sys
sys.path.insert(0, '.')
import modules.parser_Shankarappa as pS
from modules.helper import is_nonsyn_table
from modules.alphabet import alpha
# Script
if __name__ == '__main__':
# Define the patients
patients = pS.parse_sequences(reference='HXB2')
for k, p in enumerate(patients):
if str(p) != 'p10':
continue
print p
p.filter_only_sequenced()
# Measure the allele frequencies (of all alleles at all positions)
paf = p.allele_frequencies
afs = np.asarray([paf(seqs=p.seqs_from_visit(v)) for v in p.visit])
# Reshape so that we get the site as first axis, the nucleotide as
# second, the time as third
afs = afs.swapaxes(0, 2)
from conservation_syn_nonsyn_subtypeB import codon_single_mutants_synnonsyn
def get_is_mutation(consensus):
is_mutation = np.ones((len(alpha), len(consensus)), bool)
alphal = list(alpha)
for i, a in enumerate(consensus):
is_mutation[alphal.index(a), i] = False
return is_mutation.T
# Script
if __name__ == '__main__':
# Define the patients
patients = pS.parse_sequences(exclude=['p4', 'p7', 'p8', 'p11'])
# Iterate over patients
for k, p in enumerate(patients[:1]):
p.filter_only_sequenced()
# Measure the allele frequencies (of all alleles at all positions)
paf = p.allele_frequencies
afs = np.asarray([paf(seqs=p.seqs_from_visit(v)) for v in p.visit])
# Reshape so that we get the site as first axis, the nucleotide as
# second, the time as third
afs = afs.swapaxes(0, 2)
# Eliminate gaps (whole codons are excluded to keep translation possible)
# first and translate then!
# Distance from last time point to be considered likely to reach either boundary
# before the sequencing stops (later sequences will probably stay floating).
# Note: this is not used in the plot (only for testing).
maxt = {
'Shankarappa': 40, # Months
'Bunnik': 1200
} # Days
# Script
if __name__ == '__main__':
# Define the patients (excluding problematic ones)
patientsB = pB.parse_sequences(reference='SHAPE',
exclude=bad_patients['Bunnik'])
patientsS = pS.parse_sequences(reference='SHAPE',
exclude=bad_patients['Shankarappa'])
patients = patientsB + patientsS
# Counts (overall)
# The first list of each nu0s is lost, the second fixed, the third floating
# Moreover, record the patient number they came from
counts = {
x: np.zeros((len(nu0ss), 3, len(patients)), int)
for x in classes
}
# Aggregate information from all patients
for k, p in enumerate(patients):
if VERBOSE >= 1:
stderr.write(str(p) + '\n')
z = patient.U[ind, 2]
ax.scatter(x, y, z, color=cols[i], s=60)
ax.set_xlabel('PC1', fontsize=18)
ax.set_ylabel('PC2', fontsize=18)
ax.set_zlabel('PC3', fontsize=18)
ax.set_title('Patient ' + patient.name)
if show:
ppl.show()
# Script
if __name__ == '__main__':
patients = pS.parse_sequences()
for p in patients[0:1]:
print p
p.filter_only_sequenced()
# Perform PCA on all sequences
p.PCA()
# Plot
plot3D(p)
ppl.ion()
ppl.show()
fixlost = {'syn': [[], []], 'nonsyn': [[], []]}
# Counts and times (overall)
# The first list of each nu0s is lost, the second fixed, the third floating
counts_all = {
key: [[0, 0, 0] for nu0s in nu0ss]
for key in ['syn', 'nonsyn']
}
# The first list of each nu0s is lost, the second fixed
times_all = {
key: [[[], []] for nu0s in nu0ss]
for key in ['syn', 'nonsyn']
}
# Define the patients
patients = pS.parse_sequences(exclude=bad_patients)
# Aggregate information from all patients
for k, p in enumerate(patients):
if VERBOSE >= 1:
stderr.write(str(p) + '\n')
# Filter the time points to sequenced times
p.filter_only_sequenced()
n = len(p.visit)
# Filter away gaps (conservative), keep reading frame
is_nongap = ((np.array(p.seqs) == '-').sum(axis=0) == 0)
for i in xrange(len(is_nongap)):
if not is_nongap[i]: